Office Plug Load Field Monitoring Report | December 2008 1
TM
Portland, OR | San Francisco, CA | Seattle, WA | Durango, CO
Making a World of Difference
Office Plug Load Field Monitoring Report
Prepared by:
Laura Moorefield, Brooke Frazer and Paul Bendt, Ph.D.
Ecos
1199 Main Avenue, Suite 242
Durango, CO 81301
970 259-6801
December 2008
Office Plug Load Field Monitoring Report | December 2008 i
Table of Contents
Table of Figures ................................................................................................... ii
Table of Figures ................................................................................................... ii
Table of Charts .................................................................................................... ii
Abstract ................................................................................................................ 1
Introduction ......................................................................................................... 2
Study Methodology ............................................................................................. 4
Plug Load Device Selection ................................................................................................. 4
Participant Selection ............................................................................................................ 4
Recruitment ........................................................................................................................... 5
Data Collection ...................................................................................................................... 5 Meter Selection .................................................................................................................................... 5 Phase 1 Data Collection ...................................................................................................................... 6 Phase 2 Data Collection ...................................................................................................................... 7
Data Analysis .......................................................................................................................10 Processing the Meter Data ................................................................................................................ 11 Averaging and Scaling the Meter Data ............................................................................................ 12
Results .................................................................................................................................13
Business Type Results ........................................................................................................13
Analysis of Plug Load Product Categories ........................................................................17
Product Level Results .........................................................................................................18 Computers and Monitors .................................................................................................................. 18 Office Electronics .............................................................................................................................. 24 Miscellaneous .................................................................................................................................... 29
Comparing Results with Other Studies ..............................................................................31
Conclusions and Next Steps ............................................................................ 32
Consumer Implications .......................................................................................................34
Utility and Policy Implications ............................................................................................34
Future Research ..................................................................................................................34
Reference List .................................................................................................... 36
Appendices ........................................................................................................ 38
Methodology Design Matrix ................................................................................................38
Study Participant Data .........................................................................................................39
Device List with Metering Prioritization .............................................................................41
Determination of Power States ...........................................................................................46
Average Power by Mode .....................................................................................................47
Average Annual Energy Year per Mode per Device ..........................................................49
Miscellaneous Category Duty Cycle Data ..........................................................................51
Office Plug Load Field Monitoring Report | December 2008 ii
Table of Figures
Figure 1. California's Office Electricity Consumption .................................................................................... 2 Figure 2. Watts Up Pro ES Meter ................................................................................................................. 6 Figure 3. Sample Installations of Meter ........................................................................................................ 7 Figure 4. Meter Connected to Computer for Downloading Data ................................................................... 8 Figure 5. Desktop Computer Meter Data from Two-Week Metering Period ................................................. 8 Figure 6. Desktop Computer Power Data Sorted into Power States .......................................................... 12 Figure 7. 24-Hour Snapshot of a Metered Site ........................................................................................... 14 Figure 8. Annual Energy Use per Square Foot ........................................................................................... 15 Figure 9. Annual Energy Use per Full-Time Employee .............................................................................. 16 Figure 10. Office Plug Load Categories with Percentages of Total Energy Use ........................................ 17 Figure 11. Percentage of Office Energy Consumed by Product Types ...................................................... 18 Figure 12. Computer and Monitor Power Demand by Mode ...................................................................... 20 Figure 13. Computer and Monitor Energy Annual Energy Use per Device ................................................ 22 Figure 14. Average Hourly Notebook Computer Energy Use (Weekday) .................................................. 23 Figure 15. Average Hourly Desktop Computer Energy Use (Weekday) .................................................... 23 Figure 16. Imaging Equipment Power Demand by Mode ........................................................................... 24 Figure 17. Computer Peripheral Power Demand by Mode ......................................................................... 25 Figure 18. Imaging Equipment Annual Energy Use per Device ................................................................. 27 Figure 19. Computer Peripheral Annual Energy User per Device .............................................................. 28 Figure 20. Average Hourly Laser Printer Energy Use (Weekday) .............................................................. 29 Figure 21. Miscellaneous Equipment Power Demand by Mode ................................................................. 30 Figure 22. Miscellaneous Office Equipment Annual Energy Use by Mode ................................................ 31
Table of Charts
Table 1. Meter Allocation According to Prioritization .................................................................................... 9 Table 2. Computer and Monitor Duty Cycles .............................................................................................. 21 Table 3. Imaging Equipment Duty Cycles ................................................................................................... 26 Table 4. Computer Peripheral Duty Cycles ................................................................................................ 26 Table 5. Miscellaneous Equipment Duty Cycles ......................................................................................... 30 Table 6. Comparison of Selected Findings with Previous Research .......................................................... 32 Table 7. Comparison of Plug Load and Hard-Wired Lighting Energy Use ................................................. 33 Table 8. Study Design Prioritization ............................................................................................................ 38 Table 9. Study Participant Data .................................................................................................................. 39 Table 10. Device Prioritization with Number of Devices Counted and Metered ......................................... 41 Table 11. Average Power Use by Mode ..................................................................................................... 47 Table 12. Average Annual Energy Year per Mode per Device ................................................................... 49 Table 13. Miscellaneous Category Duty Cycle Data .................................................................................. 51
Office Plug Load Field Monitoring Report | December 2008 1
Abstract
Office equipment and other miscellaneous plug loads consume more than 20 percent of electricity in
California’s offices. However, energy use per device and device usage patterns are not well documented
for many of these plug loads. Previous research to understand the energy use of office products relied on
lab measurements of power drawn in various modes of operation combined with educated guesses about
the ways products are actually used.
In 2007 and 2008, Ecos Consulting and RLW Analytics conducted a plug load field monitoring study in
commercial offices in California. Researchers visited 47 offices and compiled an inventory of all plug load
devices found at each of the sites. The research team then installed plug load meters on a subset of
devices in 25 of these offices. The meter files consisted of two weeks of data at one-minute intervals for
each metered device. Researchers recorded power, current, voltage, and power factor with real-time time
stamps. In total, the team inventoried nearly 7,000 plug load devices and collected meter data from 470
plug load devices. This is the first study to actually measure how products are used in the office
environment.
While the scale of our study was not large enough to be statistically valid for all of California, the findings
provide detailed insights into how many and what types of plug loads are found in California’s offices, how
these devices operate in their everyday office settings, and how much energy they consume.
Among office plug loads, computers and monitors accounted for the largest share of energy in the office
plug loads study. Office electronics such as printers, faxes, multifunction devices, and computer speakers
accounted for 17 percent of plug load energy use. Miscellaneous devices such as portable lighting,
telephones, and coffee makers made up the remaining 17 percent. Much of this energy is consumed on
nights and weekends, when no one is working in these offices. In total, researchers estimated that
California’s office plug loads consume more than 3,000 GWh annually, costing business owners over
$400 million each year. The associated carbon dioxide emissions of these plug loads is more than
700,000 metric tons annually—equivalent to the carbon dioxide emissions of 140,000 cars during one
year.
As the state strives to meet the California Public Utilities Commission’s ―Big, Bold Energy Efficiency
Strategies,‖ which include a net zero energy mandate for all new commercial construction by 2030,
California will need to exploit every opportunity for office plug load energy reduction. These energy-
reduction opportunities include:
Aggressive consumer education on the energy use of office electronics
Promotion of office electronics whose power management features cannot be disabled
Promotion of highly efficient products and of highly efficient power supplies
Use of ―smart‖ plug strips and other automatic controls
Consideration of office electronics in Title 20
Consideration of switched outlets in Title 24
Office Plug Load Field Monitoring Report | December 2008 2
Introduction
Energy efficiency efforts designed for commercial buildings have traditionally targeted the building
envelope; heating, ventilation, and air conditioning (HVAC) systems; and hard-wired lighting. Recently,
however, the energy impact of the plugged-in equipment in these offices has garnered attention as a
result of the growing requirement for more electronic products that are faster and more robust than their
predecessors.
In the Annual Energy Outlook 2008, the Energy Information Administration (EIA) classifies office
equipment and personal computers as two of the three ―fastest growing [electrical] end uses‖ (p. 59). (The
third end use referenced is televisions.) In addition, the same report states that the ―increased penetration
of computers, electronics, appliances, and office equipment‖ is one of the significant ―factors that
influence growth in CO2 emissions‖ (p. 86).
Similarly, the most recent California Commercial End Use Survey (CEUS) also highlights office plug
loads. According to this study, office equipment accounts for 18 percent of electricity in California’s small
and large offices, making it the third-largest end use behind HVAC and lighting. The CEUS miscellaneous
category includes other plug loads not specified elsewhere. Separately, this category accounts for 5
percent of small and large office electricity use (see Figure 1). Findings from these two important studies
highlight the urgency of addressing energy-reduction opportunities in office plug loads. As improvements
are made to HVAC and lighting efficiency through Title 24, office plug loads, if not addressed, will account
for an even larger share of commercial electricity consumption.
Figure 1. California's Office Electricity Consumption
Credit: California Energy Commission, 2006
Miscellaneous
5%
Water Heating,
Refrigeration and
Cooking
5%
Compressors,
Motors and Process
4%
Office
Equipment
18%
HVAC
36%
Lighting
32%
Office Plug Load Field Monitoring Report | December 2008 3
Previous research to understand the energy use of office products relied on lab measurements of power
drawn in various modes of operation combined with educated guesses about the way the product was
actually used in the field. This effort is the first study to actually measure how products are used in the
office environment. Researchers visited 47 offices in California, inventoried and categorized all of the plug
loads found, and then metered a subset of these devices for two weeks at one-minute intervals. In total,
the research team inventoried nearly 7,000 plug load devices and collected 48 million metered data
points from 470 plug load devices. The data reveal not only how much power is being drawn when office
devices are on, off, or in standby, but also how many hours per day each device spends ―on‖ and in its
various low-power modes.
The primary goal of this research is not only to characterize office plug loads, but also to use the findings
to recommend policy priorities for the Energy Commission, electric utilities, and other interested
organizations so that growth of future commercial energy use can be reduced through voluntary market-
based programs and energy efficiency regulations. As the state strives to meet the California Public
Utilities Commission’s ―Big, Bold Energy Efficiency Strategies,‖ which include a net zero energy mandate
for all new commercial construction by 2030, California will need to exploit every opportunity for office
plug load energy reduction.
This report begins with a review of the study methodology in section 2, including scope and methodology
of product measurement, office participant selection and recruitment, and analysis methods. The results,
detailed in section 3, compare energy-use profiles of business types, overall product categories, and
individual products. Section 4 summarizes research and policy implications of the research.
Office Plug Load Field Monitoring Report | December 2008 4
Study Methodology
This report builds upon the methodology of the residential plug load field monitoring study conducted in
2006 by Ecos Consulting and RLW Analytics on behalf of the California Energy Commission’s Public
Interest Energy Research (PIER) division. (See Porter et al. 2006 for details.)
Plug Load Device Selection
Because the goal of the investigation is to identify and to prioritize the electronic end uses in commercial
buildings that represent the best opportunities for reducing energy consumption, researchers considered
only electrically powered office products that have not yet been thoroughly investigated and researched.
Previous estimates and measurements of total stock and energy use enabled the research team to focus
further by placing higher measurement priority on plug load products that we suspected had high overall
energy use and lower measurement priority on plug loads with low overall energy use. The scope does
not include large appliances (white goods), HVAC, or other hard-wired loads such as lighting or GFCI
outlets. A full list of products covered in the scope can be found in the Appendix, section 0. The product
list was largely based on the taxonomy developed by Nordman and Sanchez (2006) in Electronics Come
of Age: A Taxonomy for Miscellaneous and Low Power Products.
Participant Selection
A secondary goal of this research is to develop a sound methodology for conducting field measurement of
plug loads in a commercial setting. Researchers are pioneering the practice of installing meters on a wide
variety of electronic office products and use these data to characterize their energy use, but they must
deal with difficult-to-anticipate challenges in participant recruitment, data collection, and analysis. For
these reasons and because of the cost associated with the extended travel needed for a statistically valid
sample, researchers elected not to undertake a sample statistically mapped to California as a whole.
Instead, they attempted to diversify the participant sample as much as possible so that the data
represented the range of possible California plug load energy use profiles.
The design of the participant sample considers three variables that are likely to have the most significant
impact on plug load energy use profiles:
Geographic trends. The sample includes both rural (Sonoma County) and urban (San Diego
County) counties.1 Selecting these counties enabled researchers to collect data from Southern
and Northern California as well as from different utility territories.
Office type. The three offices types represented in the sample are 1) legal, accounting, and tax
services; 2) architectural and engineering; and 3) computer systems design. These reflect a
broad range of business types that are unlikely to have confidentiality concerns associated with a
field team on premises.
Office size. Both small (up to nine employees) and large (10 or more employees) office sizes for
each office type and county are included.
A total of 47 office participants, evenly distributed by the county, office type, and size variables outlined
above were selected for site visits.
1. These counties also have different climates, but because plug load usage is likely to be independent of climate, this
is not necessarily significant.
Office Plug Load Field Monitoring Report | December 2008 5
Recruitment
Researchers used a series of steps to recruit offices to participate in the study. First, letters were sent to
potential participants describing the study goals and options for participation. A follow-up phone call to
each potential participant enabled the research team to confirm the participant fit within the sampling plan
and met other logistical criteria. Once deemed eligible, participants were offered a monetary incentive in
exchange for their involvement in the study. In general, researchers offered larger incentives to
participants with large offices and offices where the team expected to be on site for relatively long
periods. Survey Sampling International, an agency specializing in providing ―call-lists‖ for sales
companies and research firms, provided office addresses and phone numbers.
Data Collection
Data collection occurred in two phases. During Phase 1, researchers conducted on-site walk-throughs of
22 office spaces to create an inventory of all plug load devices found in those spaces. Then the research
team reviewed device inventories and prioritized devices for metering in Phase 2. Phase 2 visits occurred
at 25 additional sites and consisted of the same kind of on-site walk-throughs as those of Phase 1. Based
on the product prioritization that occurred after Phase 1, researchers installed individual plug load meters
on a subset of the inventoried products and left the meters for two weeks to record time-series data. The
team then analyzed the data to determine the time each product spent in each operating mode, the
average power in each operating mode, and the estimated overall energy use for each product.
Aggregating the data enabled the team to make recommendations on devices that are ripe for energy
efficiency policy approaches and additional research.
Meter Selection
Researchers used a total of 120 Watts Up Pro ES meters to gather data on power demand, power factor
and time of use for individual plug load devices. Prior to selecting this meter, the team evaluated several
meters in the lab for comparison. In addition, the Electric Power Research Institute (EPRI) conducted a
meter evaluation in its lab to inform researchers’ selection. Ultimately, researchers selected the Watts Up
Pro ES meter because it had the following features:
Ability to record and store data from a single plug load device rather than a circuit or electrical
panel
Ability to record data at one- or two-minute intervals
Ability to store data for multiple weeks
Ability to record true r.m.s. watts, r.m.s. volts, r.m.s. amps, volt-amps, power factor, and add a
date and time stamp for each interval recording
User-friendly character (provided clear documentation, was easily programmed, and downloaded
data quickly and accurately)
Ability to record entire range of power for office plug load devices (<1 watt up to 1,800 watts) with
same model meter
Sufficient level of accuracy in specifications for this study. Meters evaluated in our laboratory met
the stated specifications.
Desired price point
Timely consumer availability (in 2007)
Office Plug Load Field Monitoring Report | December 2008 6
Researchers programmed the meters to record watts, volts, amps, volt-amps, power factor, and
maximum wattage at one-minute intervals. With these settings, the team was able to record a maximum
of 23,752 readings, or 16.5 days worth of one-minute interval data. The desired file length was 20,160
readings at one-minute intervals — exactly two weeks.
Figure 2. Watts Up Pro ES Meter
Photo credit: www.WattsUpMeters.com
Phase 1 Data Collection
During Phase 1 site visits, researchers created plug load inventories of 22 offices. Two field surveyors
visited each site. They first met with the site contact (often this was the IT or facilities manager) to answer
questions about the study, explain the survey procedure, ask the site contact questions about the facility
and the business’s environmental procurement practices, and get sign-offs on an Entrance Agreement
(required) and billing data release (optional). Field surveyors then walked through the office space
together and created an inventory of every plug load device. This inventory included device name, device
location (private office, production center, etc.), and type of power supply (internal, external, none, or
unknown). Surveyors recorded screen size for monitors and televisions, and number or U-bolts along with
the rack height for servers. If any device was unplugged or inaccessible for metering, the surveyor noted
that as well. All information was recorded on site in a database template stored on a tablet PC. Once off
site, surveyors could then upload the site data to a central database.
Office Plug Load Field Monitoring Report | December 2008 7
Phase 2 Data Collection
For Phase 2 sites, surveyors followed all of the above procedures to create the same plug load
inventories of 25 new sites. Next, the research team selected a subset of the inventoried devices for time-
series metering.
After all devices were logged into the database, surveyors determined the number of meters to be
allocated to that site using the metrics explained in the meter allocation section below. They then entered
the meter number into the database where a computer program assigned each meter to a specific device
type based on the metering prioritization. The computer program selected device types to be metered, but
not a specific device in a specific location. For example, the program may indicate that a desktop
computer should be metered, but not that the desktop computer at the reception desk should be metered.
This enabled the field surveyors to use professional judgment when selecting which devices to meter. In
no case did researchers use a meter to record data from multiple devices connected to a plug strip. If a
device selected for metering was plugged into a plug strip, the meter was installed between the device
and the plug strip. See the photographs in Figure 3 for examples of how meters were installed in the field.
1. No plug strip: Device plugged into meter;
meter plugged into wall outlet
2. Plug strip present: Device plugged into
meter, meter plugged into plug strip, plug
strip plugged into wall outlet.
Figure 3. Sample Installations of Meter
Once all the meters had been installed at the site, surveyors arranged a time for meter pickup with the
site contact. Meters were left on site 14 days, after which time surveyors returned to remove them. Off
site, surveyors downloaded the meter files to their computers, and then uploaded all meter files to the
central database. Figure 4 illustrates a meter connected to a computer for downloading data. Figure 5
depicts an example of two-week power data recorded from a desktop computer.
Office Plug Load Field Monitoring Report | December 2008 8
Figure 4. Meter Connected to Computer for Downloading Data
Figure 5. Desktop Computer Meter Data from Two-Week Metering Period
Metering Prioritization
Since the purpose of the study was to gather detailed information on office electronics, the research team
wanted to meter as many traditional office electronics (e.g., computers, printers, etc.) as possible while
collecting meter data on a wide variety of all plug load devices encountered. However, because it was not
feasible to meter every device at every site, the team needed to develop a prioritization system for
product metering. To do so, researchers determined three main areas of interest and categorized all of
the devices in the product taxonomy accordingly. The three main areas of interest were:
1. Devices whose cumulative energy use is believed to be high
2. Devices about which little is known (either duty cycle or energy consumption)
0
10
20
30
40
50
60
70
80
90
100
0 2880 5760 8640 11520 14400 17280 20160
1-Minute Intervals
Po
wer
(W)
Office Plug Load Field Monitoring Report | December 2008 9
3. Devices whose energy consumption could be reduced through use of automatic controls
The team assigned all of the products in our product taxonomy to one or more of the above categories,
enabling systematic sorting of devices into high, medium, low, or do not meter categories to ensure
capture of meter data from a targeted subset of encountered plug load devices.
While researchers decided some priorities on a case-by-case basis, in general, high-priority devices were
those expected to be among the highest energy users, or which fell into more than one of the above
categories. Devices that fell into the second or third areas of interest were ranked as medium priority.
Low-priority devices were those that piqued some interest but did not fall solidly into the top three areas.
Devices labeled do not meter were those that were very well understood and/or outside the scope of this
study (e.g., white goods), hard to meter accurately (e.g., laptop docking station), or rejected due to
difficulty or liability (e.g., servers). See Appendix, section 0, for device prioritization.
Meter Allocation
The goal was to have the majority of meters installed on high-priority devices, with medium- and lower-
priority devices receiving fewer meters. Researchers used the following distribution to allocate meters at
each site:
Table 1. Meter Allocation According to Prioritization
Priority for Metering Meter Allocation per Site
High 60%
Medium 30%
Low 10%
Do Not Meter 0%
The team allocated meters to each site proportionally based on square footage and/or number of full-time
employees. This enabled metering of multiple sites of different sizes during the same period. Researchers
installed a minimum of 10 meters and no more that 40 meters at each site, first developing two metrics to
determine the specific number of meters between 10 and 40. In the first metric, researchers divided the
site square footage by 100 to arrive at the number of meters to be installed at that site. However, this
metric was not always the best predictor of the number of plug loads at every site. For that reason, and to
allow the field surveyors to exercise professional judgment at the sites, researchers developed an
alternate metric—three meters for every full-time employee—to use if the first metric did not seem to yield
an appropriate number of meters for a specific site.
Examples for determining an appropriate number of meters are as follows:
Example for a 1,400-square-foot site: 1,400 divided by 100 = 14, so the site would get 14 meters.
Example for an 800-square-foot site: 800 divided by 100 = 8. The minimum number of meters to
be allocated to a site was 10, so this site would get 10 meters.
Example for a site with eight full-time employees: 8 multiplied by 3 = 24. This site would be
allocated 24 meters.
Billing Data Collection
Twenty-one sites — 10 sites in Pacific Gas & Electric’s (PG&E) territory and 11 sites in San Diego Gas &
Electric’s (SDG&E) territory — allowed review of their electricity billing data from February 2007 through
Office Plug Load Field Monitoring Report | December 2008 10
February 2008. In some cases, businesses that were tenants rather than owner-occupants did not have
the authority to release their billing data. In other cases, the billing data covered an entire office complex,
and billing data for the surveyed site could not be separated out.
Researchers asked every site contact who was willing (and able) to release billing data which equipment
and office areas their electric bill covered. Because of the various ways commercial office buildings are
metered, and the variety of arrangements that building tenants have with landlords, billing data did not
necessarily cover the same portion of the office space that researchers surveyed; however, the research
team collected billing data when sites were willing to get a general sense of the scale of plug loads within
overall electricity consumption.
Typical Site Visits
For Phase 1, surveyors were typically at each site for one and a half hours. The smallest sites required
only one hour; the largest took up to three hours. A typical Phase 2 visit took one-half to one hour longer
than a Phase 1 visit to a site of similar size. The meter removal visit took one-half to one hour per site.
During the course of the study, no equipment at participating sites was damaged, and to researchers’
knowledge, only one meter was disabled by a study participant.
Data Analysis
As noted earlier in this report, the scale of this study does not allow its findings to be statistically
significant for California. However, the in-depth case study approach provides a wealth of new data on
how many and what types of plug load devices are found in California’s offices and how these devices
operate (and are operated by users) in their everyday settings. The study’s findings can be used to
characterize plug load energy use in offices across California (but not to predict it precisely), to identify
products that warrant further research, and to shape future strategies for office equipment energy
savings.
Researchers established five main goals for our data analysis methodology:
1. Site plug load characterization. Determine the number of and types of plug load devices at each of
the 47 sites in the study.
2. Average power demand by mode. Determine the average power demand of each device type (e.g.,
laptop computer, inkjet printer) according to the mode of operation.
3. Duty cycles. Determine the percentage of time each device type spent in each of its operational
modes during the two-week metering period. Use these percentages to predict annual duty cycles.
4. Energy consumption. Determine total energy consumed in each mode for each device type during
the two-week metering period. Scale those findings up to predict the energy consumption by mode
annually.
5. Time of use. Evaluate the real-time energy consumption of selected metered devices.2
The findings from the data analysis process, as well as lessons learned, will help the efficiency research
community move closer to better understanding the energy impacts of plug load devices as well as how
to study this topic effectively.
2. Load curve data collected during this study will be evaluated in follow-on Grid Impacts Assessment research to be completed in 2009.
Office Plug Load Field Monitoring Report | December 2008 11
Processing the Meter Data
The first step in the meter data analysis was to review files for errors. File errors can occur for a variety of
reasons in field studies, including meter malfunction, improper meter setup or data download, and
interference in the field. Researchers reviewed the 470 time-series meter files for file length and plausible
power and voltage readings. In total, 40 files, or 8.5 percent, were eliminated. Fifteen of the eliminated
files had less than one week’s worth of data; 24 had unreasonable readings for power and/or voltage; and
one file was an exact duplicate of another.
A small number of meter files were slightly shorter than two weeks, and a few were slightly longer.
However, after the long files were trimmed not to exceed the two-week metering period, the product files
ranged from 99.2 percent to 100 percent of the desired file length. Researchers then wrote and refined a
program in Fortran for analysis of the meter data. The basic intent of the analysis was to determine for
each device metered:
The amount of time spent in each operating mode
The power demand in each of those modes
The energy consumed in each mode
One of the challenges of the analysis was that there was not enough information to determine the
operating ―mode.‖ The problem can be illustrated by the following typical definition of a ―sleep‖ mode from
the ENERGY STAR® Requirements for Computer Monitors v. 4.1:
―The reduced power state that the computer monitor enters after receiving instructions from a
computer or via other functions. A blank screen and reduction in power consumption characterize
this mode. The computer monitor returns to On Mode with full operational capability upon sensing
a request from a user/computer (e.g., user moves the mouse or presses a key on the keyboard)‖
(p.4).
In this case, the mode is defined by two quantities: power consumption and functionality. The meters
recorded power consumption but not functionality, so from the outset, the research team had only half of
the needed information.
To get around this, researchers first determined not the operating mode, but the power state. For the
purposes of this report, a product is considered to be in a power state when it spends a significant and
continuous period with its power consumption in a narrow range. An operating mode can include one or
more power states, and may also include fluctuating power levels that are not considered power states.
Fluctuating power levels are especially common in active mode. See Appendix, section 0, for a detailed
description of the Fortran program used to sort meter data into power states.
Office Plug Load Field Monitoring Report | December 2008 12
Figure 6. Desktop Computer Power Data Sorted into Power States
The final step was to collect statistics on each power state and to assign each state to an operating
mode. An analyst, who could include knowledge of the product’s typical operation and power
consumption patterns, made these assignments. Researchers used the following operational modes:
Disconnected. No power recorded. Disconnect mode means that the device is not drawing any
power. This could occur when a device was unplugged, turned off with a hard switch, or turned off
via a surge protector or plug strip.
Standby. Minimum observed steady power mode that was in the range of previously measured
standby power values for the product type.
Sleep. Steady low power mode between standby and idle. Power states were assigned to sleep
mode only after idle and standby states were identified. This mode assignment was used only for
product types where power-saving features are a know part of the products’ design.
Idle. Steady mode that falls below ―active‖ mode. A product operates here when it is prepared to
perform its intended function, but is not doing so.
Active. Device is performing its intended function. In some cases, products may have more than
one intended function and therefore a wide range of active mode power. For example, a
multifunction device demands different power levels for scanning, printing, and copying; however,
when it is performing any of these functions, the device is in active mode.
Averaging and Scaling the Meter Data
After sorting the meter data for each device into operational modes, the research team used the sorted
data to calculate the following for each metered product:
Average power demand by mode
Total time spent in each mode during the two-week metering period
Total energy consumed in each mode during the two-week metering period
0
10
20
30
40
50
60
70
80
90
100
4000 4500 5000 5500 6000
1-Minute Intervals
Po
we
r (W
)
Power State 1 Power State 2
Office Plug Load Field Monitoring Report | December 2008 13
The above findings were used to develop the following summary information for each device type (e.g.,
inkjet printer, notebook computer):
Average power demand by mode
Average percent of time spent in each observed mode
Average device energy consumption per mode for each device type
Next, researchers used the average device energy consumption findings from the two-week metering
period to predict the average annual energy consumption per mode for each device type. Scaling the data
from two weeks to one year was relatively straightforward for two reasons. The first is that the meter files
were quite uniform in length. Files that were too long were trimmed to be exactly 20,160 intervals (or
minutes, for a total of 14 days) long. Of the few files that were shorter than 20,160 intervals, none were
more than 0.8 percent short. The second reason is that, with the exception of space heaters and possibly
portable fans, the research team did not expect seasonal variations in office plug load energy use.
Finally, researchers multiplied the average annual energy for each device type by the total number of
those devices inventoried during the study. This enabled an understanding not only of the energy impact
of individual device types, but also the cumulative energy impacts of all devices in our study. Energy
findings reported in the next section are derived from metered devices and scaled up to all inventoried
devices. For example, if researchers metered five telephones but inventoried 80, the cumulative energy
impact of telephones would be the average of the annual energy use of the five metered phones
multiplied by 80. If researchers inventoried six answering machines but did not meter any, the findings
section would not report any energy use for answering machines. See Appendix, section 0, for a
complete list of all metered and inventoried items in this study.
Results
As noted in the methodology section, the sample size was not large enough to be statistically valid for
California. Readers should view the findings as characterizations of the sites studied rather than as a
representative sample of all of California’s offices. High-level results are in alignment with previous
research on commercial plug loads and therefore are very likely to be indicative of plug load energy use
in most offices. The following section discusses researchers’ findings about the businesses surveyed as
well as the plug load devices found in them. Finally, the report presents a comparison of the research
team’s findings with previous research.
Business Type Results
In total, researchers visited 47 sites. Almost half of these had fewer than 10 full-time employees;
the rest ranged in size from 10 to 275 full-time employees. The square footage of the sites visited
ranged from 350 to 38,000 square feet.3 Researchers inventoried a total of 6,943 plug load
devices and had usable two-week meter files from a subset of 430 of the inventoried devices. In
addition, the team collected one year of electricity billing data from 21 offices as well as limited
qualitative information on the purchasing and IT practices at each business. See Appendix,
section 0, for a complete listing business types, location, square footage, and number of full-time
employees for each site visited.
3. In the CEUS report (2008, p.8, Table E-1) small sites are <30,000 square feet and large sites are ≥ 30,000 square
feet. All except one of the sites were small according to this CEUS classification.
Office Plug Load Field Monitoring Report | December 2008 14
Figure 7. 24-Hour Snapshot of a Metered Site
Energy Findings
On average, plug loads consumed about 30 percent of total office electricity in the sites that participated
in the study. This was calculated by dividing the average annual plug load energy per square foot by the
annual energy reported in the electricity billing data collected from the participating sites. Because
researchers did not receive billing data from all of the surveyed businesses and did not have precise
information on what building systems the utility bills covered, the annual energy reported in the billing
data may not include heating, cooling, and/or lighting in some cases. Therefore, a 30 percent estimated
share of plug load energy use is likely to be somewhat high. However, the findings are similar to those
reported in the most recent Commercial End Use Survey (CEUS) report, where researchers concluded
that plug loads consume approximately 23 percent of total electricity usage for small and large offices
(Itron Inc. 2006).
Within the Ecos study, the plug load energy use documented translates to an average of 3.3 kWh per
year per square foot. The plug load energy density findings align well with the findings in the most recent
CEUS report. Researchers found that office electronics consumed 2.19 kWh per year per square foot in
small offices (<30,000 square feet) and 3.58 kWh per year per square foot for large offices (≥ 30,000
square feet). Miscellaneous plug loads add an additional 0.78 and 0.58 kWh per year per square foot for
small and large offices, respectively (Itron 2006, Table E-3, p.12). The office with the lowest energy use
per square foot was a legal services business, while the office with the highest energy use per square
foot was a computer systems design business. (See Figure 8 and Figure 9.) Interestingly, the computer
systems design business with the highest energy use per square foot had one of the lowest energy
usages per full-time employee. This business occupied only 350 square feet but had 22 full-time
employees. One explanation may be that many of these employees worked from remote offices.
Wednesday, November 14th, 2007
0
500
1000
1500
2000
2500
12 AM 3 AM 6 AM 9 AM 12 PM 3 PM 6 PM 9 PM 12 AM
24 Hour Snapshot
Po
wer
(watt
s)
Desktop Computers5 Metered
24 Total
LCD Monitors5 Metered
24 Total
Inkjet Printers
1 Metered
3 Total
Laser Printers1 Metered
5 Total
CRT Monitors1 Metered
2 Total
Wide Format Printer1 Metered
1 Total
Computer
Speakers
5 Metered
15 Total
Office Plug Load Field Monitoring Report | December 2008 15
Figure 8. Annual Energy Use per Square Foot
These same plug load energy findings meant that study participants used 802 kWh per year per full-time
employee, or 766 kWh per year for all employees (full- and part-time). The U.S. Census Bureau’s 2002
California Economic Census recorded that there were 1,164,306 total employees working in the
professional, scientific, and technical services sector (the NAICS category of all Ecos study participants)
at that time. Based on the census employee data, researchers estimate that plug loads in this sector
alone consume nearly 1 billion kWh per year. Using census data for all business types that are likely to
make up the majority of California’s small and large offices (NAICS codes 51 through 56), researchers
estimate that office plug load energy use for these businesses could exceed 3 billion kWh annually.
These findings align with the 2006 CEUS study (Itron 2006) where office electronics and miscellaneous
plug loads were estimated to consume 3,374 GWh (3.37 billion kWh) per year in small and large
California offices. At $0.13 per kWh, this equates to electricity expenses for office plug loads of more than
$400 million each year.
0
5
10
15
20
25
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Businesses
An
nu
al E
ne
rgy (
kW
h)
/ ft
2
Architectural/Engineering
Legal Services
Accounting/Tax
Computer Systems Design
Sonoma
San Diego
Office Plug Load Field Monitoring Report | December 2008 16
Figure 9. Annual Energy Use per Full-Time Employee
While researchers encountered a range of plug load energy use, no decisive patterns in office energy use
emerged according to Ecos’ business categories. For example, computer systems design businesses had
by far the two highest energy densities; however, many computer systems design businesses fell right
around the average.
The surveyed offices contained an average of seven devices per employee (full- and part-time). The
survey also revealed that, on average, there are 30 plug load devices for each 1,000 square feet of office
space.4 These results indicate that approximately 30 million plug load devices may currently be in use in
California’s offices.
Energy Efficiency Measures
Researchers learned the following qualitative information about the energy saving and other ―green‖
practices of the surveyed businesses:
Four percent stated that their IT equipment is set to manage power use — that is, to automatically
drop into a lower power state when a device is not in use.
Seven percent said that they discourage the use of screen savers; the same percentage allows
employees to change computer power management settings themselves.
4. The 2006 CEUS study (Itron, page 8, Table E-1) reported that California’s small and large offices total 1,022,013,000 square feet.
0
1000
2000
3000
4000
5000
6000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Businesses
An
nu
al E
ne
rgy (
kW
h)
/ F
ull T
ime
Em
plo
ye
e
Architectural/Engineering
Legal Services
Accounting/Tax
Computer Systems Design
Sonoma
San Diego
Office Plug Load Field Monitoring Report | December 2008 17
Twelve of the businesses surveyed stated that they have sustainable energy procurement
guidelines in place.
Five of these businesses specified that their sustainable energy procurement practices include
the purchase of ENERGY STAR® computers and monitors.
Seventy-five percent of the businesses surveyed said that they participated in some
environmental stewardship activities with 97 percent of these describing environmental
stewardship as only recycling.
One business mentioned using San Diego Gas &Electric’s energy conservation program.
Another business produced its own green power with photovoltaic panels.
While the majority of businesses in our study stated that they strive for environmental stewardship, to
almost all of them, this consisted mainly of recycling. In addition, only five businesses mentioned
procurement of ENERGY STAR computers and monitors. These findings indicate that there is still a need
for consumer education on the energy impacts of office plug loads. In addition, because power
management on IT equipment typically was not tightly regulated, another energy reduction approach in
offices would simply be to activate automatic power management settings.
Analysis of Plug Load Product Categories
To assess the overall impacts of the varieties of plug loads found in offices, researchers multiplied the
average annual energy consumption of each device type metered (e.g., inkjet printers, LCD monitors,
etc.) with the total number of each device type inventoried in all offices in this study. This enabled
researchers to expand upon the metered data findings to estimate the annual energy consumption of all
plug loads encountered in the study. Researchers then categorized all of plug loads in this study into one
of three groups: Computers and Monitors, Office Electronics, and Miscellaneous Plug Loads. Figure 10
illustrates the distribution of cumulative annual energy use per product category at the 47 sites visited.
Figure 10. Office Plug Load Categories with Percentages of Total Energy Use
As expected, the Computers and Monitors category was by far the largest category, accounting for 66
percent of all plug load energy use at the offices in the study. This category includes desktop, laptop, and
Miscellaneous,
17%
Office
Electronics, 17% Computers and
Monitors, 66%
Office Plug Load Field Monitoring Report | December 2008 18
thin client computers as well as CRT and LCD monitors. This suggests that programs and policies
targeting these two most basic of office electronics stand to have the greatest energy reduction.
The other two categories, Office Electronics and Miscellaneous, each account for 17 percent of plug load
energy use. Office Electronics includes imaging equipment (e.g., printers, copiers, and multifunction
devices) as well as computer peripherals such as computer speakers, external drives, and hubs and
switches. Devices such as paper shredders, adding machines, and portable desk lamps fall into the
Miscellaneous category. Miscellaneous also includes telephony equipment and small kitchen appliances
like coffee makers and toaster ovens. (White goods such as dishwashers and refrigerators are outside
the scope of this study and therefore not included in the above categories.)
Product Level Results
This section discusses the study’s findings in detail by presenting results on average power demand by
mode, duty cycle, and estimated annual energy use on a product by product basis.
Figure 11. Percentage of Office Energy Consumed by Product Types
Computers and Monitors
Computers were the single largest office plug load end use. Their energy consumption alone accounted
for 46 percent of total office plug load energy use in the study. Researchers collected meter data on 61
desktop computers, 20 notebooks, and six thin client computers.
Power Demand by Mode
The low power modes (sleep and standby) of desktop and notebook computers in the study were similar,
typically less than 3 watts. All of the computers metered demonstrated low standby power values (0.9 to
2.6 watts), indicating that efficiency standards targeting standby mode have effectively lowered standby
power in computers.
Telephony, 2% Business
Equipment, 14%
Computer
Peripherals, 6%
A / V, 1%
Computers, 46%
Monitors, 20%
Imaging, 11%
Office Plug Load Field Monitoring Report | December 2008 19
Active mode power was also similar for these two technologies. The gap between notebooks and
desktops has significantly decreased from Ecos’ previous residential plug load (Porter, et al. 2006) study
to only a 4.2 watt difference. Desktop computers averaged 79 watts in active mode while notebooks
averaged 75 watts in active mode. However, when comparing desktop and notebook computer power, it
is important to keep in mind that desktop computer power demand does not include power attributed to its
display while notebook power demand is likely to include power attributed to the display. Laptop power
demand may also include power drawn to charge the battery.
Idle mode power in both desktop and notebook computers warrants some discussion. Average desktop
idle power was 46 watts, while notebook idle power averaged 30 watts. Thin client idle power was 31
watts on average. ENERGY STAR version 4.0 maximum idle power levels range from 50 watts to 95
watts for desktops and 14 watts to 22 watts for notebooks depending on the class of the computer.
Study findings for average idle power were somewhat lower than expected given that the Tier 1 ENERGY
STAR program requirements for computers became effective July 20, 2007. In addition, while the 80
PLUS5 program for computer internal power supplies has very likely contributed to industry wide
improvements in computer power supply efficiency, the 80 PLUS program manager at Ecos Consulting
estimates a 3 percent to 5 percent penetration rate for 80 PLUS power supplies in California (personal
communication, Rasmussen, October 13, 2008). While this is a significant achievement, it cannot solely
account for the study’s low idle power findings. The most plausible explanation is that, as previously
discussed, the research methodology used in this study to sort metered power data into power states
used a somewhat different approach to identify idle mode than did previous research, including ENERGY
STAR’s. ENERGY STAR defines idle mode as a set of functions whereas in this study, researchers
recorded only power values, not function. Therefore, what is categorized as idle mode here is likely to be
the steady, lower end of the range of power values that would be considered an idle mode as defined by
a set of functions. The discrepancy between computer idle power findings from this study and previous
research is due to different approaches of categorizing data rather than from discrepancies in the data
itself.
Nonetheless, previous PIER research indicates that idle power in computers can be reduced below the
power levels researchers identified as well as the ENERGY STAR levels. In ―How Low Can You Go: A
White Paper on Cutting Edge Commercial Desktop Computer Efficiency‖ (Beck et al. 2008), researchers
found that desktop computer idle power can be reduced to 30 watts with off-the-shelf components and to
just 19 watts with best-in-class computer components. Given this, idle mode presents a ready opportunity
for computer power reduction.
For thin client computers, note that idle power is higher than active power. The reason is that only three of
the six metered thin client computers metered exhibited an idle mode. In addition, two of these three
computers had much higher active power (44 watts each) than the other four models (15 watts, 15.7
watts, 16 watts, and 23 watts). While the high active power of the outlier models was averaged in with
lower active power of all six thin client computers, the idle mode depicted in the graph is representative of
only those three thin client computers that had an idle mode.
5. See http://www.80plus.org/ for details on the 80 PLUS program.
Office Plug Load Field Monitoring Report | December 2008 20
Figure 12. Computer and Monitor Power Demand by Mode
The power data for monitors was derived from meter files from 21 CRT monitors and 84 LCD monitors.
This is representative of the distribution of all monitors inventoried. As expected, LCD monitors drew less
power than CRT monitors in every mode, with significant differences in every mode except standby.
Within the monitor category, CRTs used 14 to 49 percent more power than LCDs in each mode. CRTs
used an average of 71 watts in active, while LCDs used an average of only 34 watts in active. The
average power for CRTs in sleep mode was 46 watts while the average power for LCDs in sleep mode
was just over 6 watts. Of the 955 monitors inventoried in the 47 offices in this study, 79 percent were
LCDs, indicating that consumers are already making the shift to the more-efficient technology.
Duty Cycles
Standby mode was the predominant mode for all computers occupying between 39 percent (for thin client computers) up to 56 percent (for notebook computers) of the two-week metering period. On average, desktop computers spent one-third of the two-week metering period in active mode. Thin client computers had a similar amount of time devoted to active mode. In contrast, notebook computers spent only 10 percent of time in active mode. Idle mode was utilized the most by thin client computers. Desktops and notebooks used idle mode less frequently at 13 percent and 6 percent of time, respectively. Sleep mode was not utilized often in any of the computers metered. Time spent in disconnect ranged widely—from 1 percent for thin clients to 26 percent for notebooks. Disconnect mode could indicate that the computer was unplugged. More likely is that the computer was powered down with a hard off switch or turned off via a plug strip. Because thin client computers are often used as small servers, their low percentage of time in disconnect makes sense.
0
10
20
30
40
50
60
70
80
90
N=61 N=20 N=6 N=21 N=84
Desktop Notebook Thin client CRT LCD
COMPUTERS MONITORS
Po
we
r (W
)Active
Idle
Sleep
Standby
Sleep = 3.2 W
Standby = 2.2 W
Office Plug Load Field Monitoring Report | December 2008 21
Table 2. Computer and Monitor Duty Cycles
Product Number Metered
Average Time in Active
Average Time in
Idle
Average Time in Sleep
Average Time in Standby
Average Time in
Disconnect
Desktop Computer
61 30% 13.4% 0.4% 50% 7.2%
Notebook Computer
20 10% 6% 2% 56% 26%
Thin client 6 29% 29% 2% 39% 1%
CRT Monitor
21 17% 2% 0.4% 48% 32.6%
LCD Monitor
84 18% 8% 2% 50% 22%
CRT and LCD monitors had very similar duty cycles. LCDs spent more time in idle mode than did CRT
monitors, but researchers found that CRT monitors were in disconnect mode for 10 percent more time
than LCD monitors. Disconnect mode for monitors is likely to indicate that the monitor is powered down
with a hard off switch.
Energy Use by Mode
Results from this study indicate that the average desktop computer consumes 407 kWh per year
compared with 96 kWh per year for the average notebook computer. These annual energy findings for
desktop computers are in alignment with previous PIER research (Beck et al. 2008). The average CRT
monitor consumes approximately 220 kWh per year, and the average LCD monitor consumes 132 kWh
per year. With both desktop computers and CRT monitors, the majority of the energy consumed annually
is attributable to active mode. In contrast, annual energy consumption is distributed more evenly across
active and idle modes for notebook computers and LCD monitors. The energy use difference between
LCD and CRT monitors in this study was not as large as we expected. One explanation is that LCD
monitors tend to be larger than CRT monitors. Also, they spent more time in idle and sleep modes than
did CRTs, and were turned off less often. LCD monitor energy use could be reduced by decreasing the
time spent in idle mode on nights and weekends.
Office Plug Load Field Monitoring Report | December 2008 22
Figure 13. Computer and Monitor Energy Annual Energy Use per Device
Time of Use
Because computers made up such a significant share of office plug load energy use, researchers
conducted an additional time of use analysis on desktop and notebook computers to better understand
how and when computers operate. Figure 14 and Figure 15 illustrate the average hourly energy use per
mode of notebook and desktop computers, respectively. The energy levels represented in these graphs is
the weekday hourly average of the 20 notebook and 61 desktop computers metered. Note how active
mode is concentrated during working hours for notebooks but continues throughout the evening for
desktops. Active energy use by computers during nonworking hours is often indicative of screen savers.
Sleep mode accounts for very little energy and is rarely used.
0
50
100
150
200
250
300
350
400
450
N=61 N=20 N=6 N=21 N=84
Desktop Notebook Thin client CRT LCD
COMPUTERS MONITORS
Annual E
nerg
y U
se (
kW
h)
active
idle
sleep
standby
Office Plug Load Field Monitoring Report | December 2008 23
Figure 14. Average Hourly Notebook Computer Energy Use (Weekday)
Figure 15. Average Hourly Desktop Computer Energy Use (Weekday)
Notebook Computer
0
5
10
15
20
25
30
12-1
am
1-2
am
2-3
am
3-4
am
4-5
am
5-6
am
6-7
am
7-8
am
8-9
am
9-1
0 a
m
10-1
1 a
m
11-1
2 a
m
12-1
pm
1-2
pm
2-3
pm
3-4
pm
4-5
pm
5-6
pm
6-7
pm
7-8
pm
8-9
pm
9-1
0 p
m
10-1
1 p
m
11-1
2 p
m
Wh
active
idle
sleep
standby
Desktop Computer
0
5
10
15
20
25
30
35
40
45
50
55
60
12-1
am
1-2
am
2-3
am
3-4
am
4-5
am
5-6
am
6-7
am
7-8
am
8-9
am
9-1
0 a
m
10-1
1 a
m
11-1
2 a
m
12-1
pm
1-2
pm
2-3
pm
3-4
pm
4-5
pm
5-6
pm
6-7
pm
7-8
pm
8-9
pm
9-1
0 p
m
10-1
1 p
m
11-1
2 p
m
Wh
active
idle
sleep
standby
Office Plug Load Field Monitoring Report | December 2008 24
Office Electronics
In this study, the office electronics category consisted of imaging equipment and computer peripherals.
Printers, fax machines, scanners, and multifunction devices were all considered imaging equipment and
made up 11 percent of plug load energy use in the offices that participated in the study. In total,
researchers metered 77 imaging devices and inventoried 232. Laser printers accounted for more than
half of all the imaging equipment in the study.
Computer peripherals metered in this study consisted mostly of computer speakers, but also included
external drives, and Ethernet and USB hubs and switches. These devices accounted for 6 percent of
office plug load energy use.
Power Demand by Mode
Most of the imaging devices metered operated in standby, idle, and active modes throughout the two-
week metering period; however, a sleep mode was also apparent in some samples (but not all) of two
device types: inkjet printers and laser multifunction devices. As expected, active mode power was higher
in laser devices than in inkjet devices. An exception was the wide format printers, which are typically
inkjet. The laser printer had the highest active power demand at 130 watts, followed by the wide format
printer active power of 87 watts and the laser multifunction devices’ average active power of 76 watts.
The process of fusing ink onto paper utilized in laser printers is energy intensive because it requires a
great deal of heat. Therefore, a laser printer has higher active power than a laser multifunction device
because the fuser is needed for every print job but is not needed for faxing and scanning.
Figure 16. Imaging Equipment Power Demand by Mode
0
20
40
60
80
100
120
140
N=18 N=3 N=33 N=13 N=7 N=1 N=2
Laser Inkjet Laser Inkjet Wide format
MULTI-FUNCTION DEVICE PRINTER DOCUMENT
SCANNER
LASER FAX
Po
we
r (W
)
Active
Idle
Sleep
Standby
Idle = 6.8 W
Sleep = 4.7 W
Standby = 2.7 W
Sleep = 5.4 W
Standby = 5.5 W
Office Plug Load Field Monitoring Report | December 2008 25
Across the range of imaging equipment, idle mode power was well below active, but on average more
than three times higher than standby. This indicates potential for power reduction and ultimately energy
savings by enabling device power management features to automatically move the device into a low
power mode instead of remaining in idle mode indefinitely.
Note that only one scanner and two fax machines were metered. While these files reveal useful data
about the metered devices, they can in no way be considered to be representative samples for these
technologies.
All of the computer peripheral devices metered had active power demands of less than 30 watts with the
exception of two samples of computer speakers. Eighteen of the 20 computer speakers metered had
active power in the range of 7 to 8 watts. This finding was consistent with the computer speaker data
recorded in the 2006 residential plug load field research (Porter et al. 2006). The two outlier computer
speakers had active power values of 78 watts each. These high active measurements were surprising but
probably recorded from computer speaker systems that often include multiple speakers and a subwoofer
all linked together in one system. While such high-power computer speakers are not likely to be found in
many office settings, they are evidently present in some. Researchers chose to separate out the higher
power readings along with those of traditional computer speakers for clarity. Standby mode was 3 watts
or less for every device in this category.
Figure 17. Computer Peripheral Power Demand by Mode
Duty Cycles
Standby mode dominated all of the imaging equipment devices metered. However, laser multifunction
devices, laser printers, inkjet printers, and wide format printers all showed time in disconnect mode.
0
5
10
15
20
25
30
35
40
N=20 N=2 N=2 N=9 N=2
Traditional High Power Ethernet USB
COMPUTER SPEAKERS EXTERNAL
DRIVE
HUB OR SWITCH
Po
we
r (W
)
Active
Idle
Sleep
Standby
70
75
High Power Computer Speakers: 73 W
Office Plug Load Field Monitoring Report | December 2008 26
These devices can have hard off switches; it appears that in several instances, these devices were
actually turned off. Laser multifunction devices and laser printers showed the highest percentage of the
two-week metering period in active mode—14 percent each. These two products along with wide format
printers had the highest percentage of time in idle mode as well. Sleep mode was very rarely used— only
the inkjet printers exhibited this mode.
Table 3. Imaging Equipment Duty Cycles
Product Number Metered
Average Time in Active
Average Time in
Idle
Average Time in Sleep
Average Time in Standby
Average Time in
Disconnect
Laser MFD 18 14% 14% 0% 66% 6%
Inkjet MFD 3 1% 2% 0% 97% 0%
Laser Printer 33 14% 17% 0% 51% 18%
Inkjet Printer 13 2% 4% 5% 68% 21%
Wide Format Printer 7 6% 34% 0% 33% 27%
Document Scanner 1 3% 0% 0% 97% 0%
Laser Fax 2 4% 0% 0% 96% 0%
In contrast, many computer peripherals spent the majority of the metering period in idle mode. Computer
speakers operated in idle for 84 percent of the time, and hubs and switches, both Ethernet and USB,
operated in idle for 55 percent of the time. Standby was the dominant mode for the two external drives
metered.
Table 4. Computer Peripheral Duty Cycles
Product Number Metered
Average Time in Active
Average Time in
Idle
Average Time in Sleep
Average Time in Standby
Average Time in
Disconnect
Traditional Computer Speakers
18 1.5% 89% 0% 4% 5.5%
High End Computer Speakers
2 30% 0% 0% 7% 63%
External Drive 2 10% 0% 4% 86% 0%
Ethernet Hub or Switch 9 16% 53% 0% 20% 11%
USB Hub or Switch 2 2% 55% 4% 39% 0%
Energy Use
Wide format printers, laser printers, and laser multifunction devices had the highest annual energy use
per device. This makes sense because these are the same devices that had the highest active and idle
power demands as well as the highest percentages of time in both of these modes. Laser printers in the
study consumed 280 kWh per year on average. Laser multifunction devices followed suit, consuming just
Office Plug Load Field Monitoring Report | December 2008 27
under 200 kWh per year. Wide format printers consumed more than 320 kWh per year; however,
because these printers are typically employed in only architectural and engineering offices, they do not
represent a typical load for many offices.
Both inkjet printers and multifunction devices use only a third of the overall energy use of their laser
counterparts. However, in the study’s product inventory, laser printers and multifunction devices
outnumbered inkjet printers and multifunction devices by 3 to 1. A simple energy savings strategy would
be to utilize inkjet technology instead of laser technology whenever possible.
With the exception of the laser printer, standby energy use was below 50 kWh per year for products in
this category. For the 33 laser printers that metered, the average annual standby energy was 88 kWh.
This energy use in standby mode alone is more than what a normally operated6 75-watt light bulb would
consume over the course of one year.
Figure 18. Imaging Equipment Annual Energy Use per Device
Low power modes account for a significant share of energy use in many computer peripherals. Much of
this energy consumption could be eliminated through the use of ―smart‖ plug strips. These devices use a
timer, load sensor, occupancy sensor, or some combination thereof to shut off power to selected devices.
In an office setting, a smart plug strip could be used to cut power to the monitor and computer peripherals
when the computer enters a sleep, standby, or disconnected mode. Alternatively, a smart plug strip with
an occupancy sensor could be programmed to power down selected devices when no occupant is
present. A timer controlled plug strip would be an effective energy reduction solution for devices that do
6. Assumes 1,000 hours of operation per year
0
50
100
150
200
250
300
350
N=18 N=3 N=33 N=13 N=7 N=1 N=2
Laser Inkjet Laser Inkjet Wide format
MULTI-FUNCTION DEVICE PRINTER DOCUMENT
SCANNER
LASER FAX
Annual E
nerg
y U
se (
kW
h)
active
idle
sleep
standby
Office Plug Load Field Monitoring Report | December 2008 28
not need to draw power at night and on weekends. Through any of these methods, significant energy
savings could be realized in most computer peripherals through responsive ―smart‖ controls.
Figure 19. Computer Peripheral Annual Energy User per Device
Time of Use
Researchers conducted a time of use study for laser printers to better understand the time of day energy
use by these devices. While the printer energy use peaks during the expected time frame, energy
consumed in all modes remained relatively high overnight. Sleep mode was not utilized for these devices.
Given this data, laser printers appear to be a ready target for after-hours energy reduction programs.
0
20
40
60
80
100
120
140
N=20 N=2 N=2 N=9 N=2
Traditional High Pow er Ethernet USB
COMPUTER SPEAKERS EXTERNAL DRIVE HUB OR SWITCH
Annual E
nerg
y U
se (
kW
h)
active
idle
sleep
standby
Office Plug Load Field Monitoring Report | December 2008 29
Figure 20. Average Hourly Laser Printer Energy Use (Weekday)
Miscellaneous
Energy use by all devices in the Miscellaneous category accounted for 17 percent of plug load energy
use for the offices in the study. The Miscellaneous category includes audio/visual equipment, telephony,
and general business equipment such as paper shredders, adding machines, portable lamps, and coffee
makers. Energy use of audio/visual equipment accounted for only 1 percent of office plug load energy use
while telephony accounted for 2 percent. General business equipment was the largest share of the
Miscellaneous category and accounted for 14 percent of office plug load energy use.
Because the Miscellaneous category comprises many disparate devices, in this section we discuss only
the devices with the highest cumulative energy consumption: coffee makers, portable lighting, and paper
shredders. Detailed findings for all devices in this category are available in the Appendix, sections 0
through 0.
Surprisingly, coffee makers were the largest energy consumer (cumulatively) in this category. Based on
two weeks of meter data, researchers estimated that these devices consume nearly 800 kWh per year
per device — the same energy that two normally used desktop computers would consume in one year.
Coffee makers in this study had an average active power demand of 464 watts. This is likely due to the
fact that, while many offices have single-pot coffee makers typically found in homes, many models in
offices are the larger, commercial variety. These coffee makers may brew two pots at one time, or brew
coffee directly into stationary containers where coffee is then dispensed through spouts. In addition,
researchers observed that while some coffee makers had an intermediate ―keep warm‖ power level, other
models simply cycled a high-power heating element on and off to keep the coffee at the appropriate
temperature throughout the day. The study’s inventory included a total of 50 coffee makers.
Portable lighting was the second-largest cumulative energy consumer in this category. Researchers
recorded meter data from 156 table lamps and 236 desk attachment lamps. Desk attachment lamps
include lamps clamped to desks or cubicle walls, or plug-in fixtures installed under office cabinets. While
researchers did not record lamp technology in the field, their standby power and power factor findings
indicate that a variety of technologies — incandescent, fluorescent tubes, compact fluorescent lights, line-
Hourly Energy Use of a Laser Printer
0
5
10
15
20
25
30
35
40
45
50
12-1
am
1-2
am
2-3
am
3-4
am
4-5
am
5-6
am
6-7
am
7-8
am
8-9
am
9-1
0 a
m
10-1
1 a
m
11-1
2 a
m
12-1
pm
1-2
pm
2-3
pm
3-4
pm
4-5
pm
5-6
pm
6-7
pm
7-8
pm
8-9
pm
9-1
0 p
m
10-1
1 p
m
11-1
2 p
m
Wh
active
idle
standby
Office Plug Load Field Monitoring Report | December 2008 30
voltage halogen, and low-voltage halogen — are all in use in portable office lighting. Annual energy for
individual lamps, both table and desk attachments — approximately 100 kWh — was relatively low;
however, the prevalence of portable lighting in offices accounts for the notable energy impact.
Electric paper shredders were the third-largest energy end use in this category. Individually, these
devices consume 168 kWh per year, and researchers recorded 60 of them in the plug load inventory. As
illustrated in Figure 22, active mode energy dominates the total energy use of these devices.
Figure 21. Miscellaneous Equipment Power Demand by Mode
Table 5. Miscellaneous Equipment Duty Cycles
Product Number Metered
Average Time in Active
Average Time in
Idle
Average Time in Sleep
Average Time in Standby
Average Time in
Disconnect
Desk Attachment Lamp 9 20% 1% 0% 9% 70%
Table Lamp 16 14.5% 1.5% 0% 4% 80%
Coffee Maker 10 25.5% 16% 0% 46% 12.5%
Shredder 4 25% 0% 0% 43% 32%
0
10
20
30
40
50
60
70
80
90
100
110
N=12 N=18 N=10 N=4
Desk Attachment Table
LAMP COFFEE MAKER SHREDDER
Po
we
r (W
)
Active idle
sleep standby
460
470
480Coffee Maker: 464 W
Office Plug Load Field Monitoring Report | December 2008 31
Figure 22. Miscellaneous Office Equipment Annual Energy Use by Mode
Comparing Results with Other Studies
As noted above, the study’s high-level findings are in alignment with the 2006 CEUS report (Itron 2006)
for share of total office energy consumed by plug loads, office plug load energy density, and total
California energy consumed by office plug loads. Because the study was not large enough to be
statistically significant, some variances from previous research in this area were expected.
The 2007 LBNL study, Space Heaters, Computers, Cell Phone Chargers: How Plugged In Are
Commercial Buildings? (Sanchez et al. 2007) reported slightly lower results than the Ecos study did. In
this report, researchers found that the plug loads accounted for 11 percent to 19 percent of total electricity
consumption in the buildings they audited. One explanation for this difference in findings is that the LBNL
sites were in San Francisco, Atlanta, and Pittsburg, whereas all of the Ecos study sites were in California.
Because California has a mild climate and many building efficiency standards, it is reasonable that the
share of office plug load energy use is higher in this state than in states with fewer regulations and
greater heating and cooling loads. In the Annual Energy Outlook (2006), the EIA found that 14 percent of
commercial electricity was attributable to PC and non-PC office equipment. This lower percentage is
expected because the Ecos study focused specifically on offices, not the entire commercial sector, where
office electronics and other plug loads are likely to be more prevalent. Similarly, the 2006 PG&E report
Consumer Electronics: Market Trends, Energy Consumption, and Program Recommendation (Chase et al 2006)
reported that in 2005, electronics represented 18 percent of small business and residential electricity
consumption in PG&E’s territory. Regarding the number of devices per square foot, LBNL reported 23
plug load devices per 1,000 square feet for all the commercial buildings it surveyed, whereas Ecos
0
100
200
300
400
500
600
700
800
900
N=12 N=18 N=10 N=4
Desk attachment Table
LAMP COFFE MAKER SHREDDER
Annual E
nerg
y U
se (
kW
h)
active
idle
sleep
standby
Office Plug Load Field Monitoring Report | December 2008 32
recorded 30 devices per 1,000 square feet. However, the average number of devices for small, medium,
and large offices in the LBNL study is 31.5 devices per 1,000 square feet (Sanchez et al. 2007).
Perhaps the most significant departure from previous research is the Ecos study’s findings on computer
energy use. Previous research may have underestimated the magnitude of computer energy use. Annual
per-product energy use was not measured in previous research, but rather was calculated based on
instantaneous power measurements and estimates of annual hours spent in each mode. Average
measured commercial desktop and notebook computer annual energy per device is 1.3 to 3.9 times
higher than these previously calculated values (see Table 6). Another explanation for the discrepancy in
computer energy use is very likely the fact that both desktops and laptops in use today are faster, more
powerful, and have more capabilities than the computers that were evaluated in previous studies.
Table 6. Comparison of Selected Findings with Previous Research
Study Annual
kWh/product
LBNL Annual kWh/product (Kawamoto et
al 2001)
LBNL Annual kWh/product (McWhinney et al 2004)
LBNL Annual kWh/product (Sanchez et al
2007)
US DOE Annual
kWh/product (Roth 2002)
Desktop Computer 407 213 n/a n/a 297
Laptop/Notebook
Computer 96 24.6 n/a n/a 32
Inkjet Printer 84 74 52 n/a 92
Laser Printer 280 283 6207 n/a 735
LCD Monitor 132 n/a n/a n/a 23
CRT Monitor 220 205 n/a n/a 306
Inkjet MFD 47 n/a 73 n/a n/a
Laser MFD 197 n/a n/a n/a n/a
Computer speakers 45 n/a n/a 74 n/a
Conclusions and Next Steps
The purpose of this study was to gain detailed insights into the plug loads currently in use in California’s
offices and to inform future energy efficiency policies, programs, and research. Researchers inventoried
how many and what kinds of plug loads are in use in 47 California offices, and recorded detailed meter
files to determine which modes products operate in, what the average power demand is for each
identified mode, and how products operate and are operated by consumers in their everyday office
settings.
7. Includes data from black-and-white and color laser printers
Office Plug Load Field Monitoring Report | December 2008 33
Furthermore, using simple, scenario-based calculations from study findings, researchers believe that in
some circumstances, energy use by office plug loads can exceed the energy use of hard-wired lighting.
Consider a hypothetical 10-foot-by-12-foot private office. Depending on the efficiency of the devices in
use and the efficiency of the hard-wired lighting, office plug loads could consume more than four times
the energy of the hard-wired office lighting.
Table 7. Comparison of Plug Load and Hard-Wired Lighting Energy Use8
Office Equipment
Active Power
(W)
Energy
(kWh)/year
Energy (kWh)
per year per ft2
Traditional
Desktop Computer 79.0 407.0 3.4
CRT Monitor 71.0 220.0 1.8
Computer Speakers 7.0 45.0 0.4
Telephone 4.8 20.0 0.2
Desk lamp: 60 W incandescent 60.0 75.0 0.6
Total 161.8 691.9 5.8
High Efficiency
Laptop 75.0 96.2 0.8
Computer Speakers 7.0 45.0 0.4
Telephone 4.8 20.0 0.2
Desk lamp: 15 W LED 15.0 18.8 0.2
Total 101.8 180.0 1.5
Hard-Wired Lighting
Active Power
(W)
Energy
(kWh)/year
Energy (kWh)
per year per ft2
Traditional
Two 4x2 Troffers, 3 standard T-12 lamps
each 178.0 445 3.7
High Efficiency
Two 4x2 Troffers, 1 efficient T-5 lamp each 70.0 175 1.5
8. These are hypothetical scenarios. Notes and assumptions:
Plug load power and energy data are findings from commercial plug load field research.
Desk lamps assumed to operate for five hours per day, five days per week, 50 weeks per year.
Hard-wired lighting assumed to operate for 10 hours per day, five days per week, 50 weeks per
year.
Hard-wired lighting equipment estimates provided by Stan Walerczyk, LC, Lighting Wizards
Office Plug Load Field Monitoring Report | December 2008 34
Consumer Implications
Office occupants can begin saving energy immediately by simply turning devices off that are not in use.
This is a no-cost energy savings strategy. Another would be to prohibit the use of screen savers that can
cause computers and monitors to operate in active mode. Researchers found that many devices were
often left to operate in active or idle mode overnight and on weekends. Offices could implement their own
awareness campaigns to educate occupants of the importance of powering down their office equipment
— just as they switch off the lights — before heading home for the evening. Alternatively, offices could
implement networked power management systems that allow IT managers to power networked devices
on and off as required for updates. Offices should also consider replacing worn-out or inefficient devices
with high-efficiency models. Notebook computers use only one-fourth of the energy consumed by
desktops. LCD monitors use far less power in active than do CRT models, but at sites surveyed, they
were often left on during nonworking hours. No matter how efficient any technology is, few devices need
to be in an active mode when no one is there to use them. Finally, while it is neither practical nor feasible
for most businesses to upgrade all equipment, offices could establish new procurement procedures with a
focus on plug load energy reduction.
Utility and Policy Implications
Utilities can look to these findings to inform new programs and policies in their territories. Rebates could
be designed for office electronics that ship with automatic controls enabled to power the device down to a
low power mode when not in use. Another program opportunity for utilities is promotion of and rebates for
―smart‖ plug strips. ―Smart‖ plug strips vary in design but typically employ some combination of load
sensors, remote controls, occupancy sensors, and timers. These inexpensive devices power down
designated plug loads when the control load is turned off by the user. The burden of responsibility to
power down electronic devices is thereby taken away from the consumer. Additional research is under
way to quantify the energy reduction potential from these devices.
Results of this study can also inform policymakers about priority products ready for new mandatory
standards or voluntary specifications. California has led the nation in mandating power supply efficiency,
but for certain products, the bar could be raised even higher through widespread implementation of power
supply efficiency programs such as ENERGY STAR, 80 PLUS, and Climate Savers. Title 20 could
address some commercial plug loads that are increasingly ready for standards consideration. Title 24
could consider a requirement for switched outlets. For example, private offices and conference rooms
could be required to have a certain percentage of their wall outlets controlled by a single switch located
near the room entrance. Automatic controls, already effectively used with hard-wired lighting, could be
required to operate some wall outlets as well.
While voluntary programs and mandatory regulations have had a vital role in improving the energy
efficiency of office plug loads, the increased reliance on office electronics coupled with a growing need for
faster, higher-power, higher quality equipment has resulted in an overall increase in plug load energy
consumption. Significant opportunities for energy savings remain untapped.
Future Research
Further research needs to be conducted to estimate the energy savings potential of all of the measures
noted above including automatic controls such as ―smart‖ plug strips, other timer or occupancy sensing
outlet controls, and widespread use of devices’ own power management settings. In addition, while
harder to quantify, future research should also explore the potential for plug load energy reduction
through consumer educations campaigns.
Office Plug Load Field Monitoring Report | December 2008 35
One important and growing end use that warrants further investigation is servers and data centers. Ecos
included these devices in the product inventory, but did not meter them due to concerns about disruption
of service. For this reason, the study’s cumulative office plug load energy use may be too low.
Finally, future research should leverage the methodologies developed during this study. A study of this
nature requires significant efforts to design the research plan, recruit participants, visit sites, install and
remove meters, transfer and review the meter files, and analyze the data. A follow-on study scaled up to
a sample size that is statistically valid for all of California’s offices could build upon our many successes
and lessons learned.
Office Plug Load Field Monitoring Report | December 2008 36
Reference List
Beck, Nathan, Peter May-Ostendorp, Chris Calwell, Baskar Vairamohan, Tom Geist. 2008. How Low Can
You Go? A White Paper on Cutting Edge Efficiency in Commercial Desktop Computers.
California Energy Commission, CEC-500-06-007.
California Public Utilities Commission. 2008. California Long‐Term Energy Efficiency Strategic Plan.
Chase, Alex, Ryan Ramos, and Ted Pope. 2006. Consumer Electronics: Market Trends, Energy Consumption,
and Program Recommendations. PG&E Application Assessment Report #0513. San Francisco,
California: Energy Solutions, for Pacific Gas and Electric
Energy Information Administration. 2006. 2003 Commercial Buildings Energy Consumption Survey:
Consumption and Expenditures Tables.
Energy Information Administration. 2008. Annual Energy Outlook 2008, with Projections to 2030.
DOE/EIA-0383 (2008).
ENERGY STAR®. n.d. ENERGY STAR® Program Requirements for Computer Monitors Eligibility
Criteria (Version 4.1). Retrieved from
http://www.energystar.gov/ia/partners/product_specs/eligibility/monitors_elig.pdf.
Itron Inc. 2006. California Commercial End-Use Survey. California Energy Commission, CEC-400-2006-
005.
Kawamoto, Kaoru, Jonathan G. Koomey, Bruce Nordman, Richard E. Brown, Mary Ann Piette, Michael
Ting, and Alan K. Meier. 2001. Electricity Used by Office Equipment and Network Equipment in
the U.S.: Detailed Report and Appendices. LBNL–45917.
McWhinney, Marla, Gregory Homan, Richard Brown, Judy Roberson, Bruce Nordman, John Busch. 2004.
Field Power Measurements of Imaging Equipment. LBNL-54202.
Nordman, Bruce, Marla Sanchez. 2006. Electronics Come of Age: A Taxonomy for Miscellaneous and
Low Power Products. LBNL-63559.
Porter, Suzanne Foster, Laura Moorefield, Peter May-Ostendorp. 2006. Final Field Research Report.
California Energy Commission, PIER Program. CEC-500-04-030.
Rasmussen, Ryan. 2008. Personal communication with 80 PLUS program manager, Ecos Consulting.
Roth, Kurt W., Fred Goldstein, and Jonathan Kleinman. 2002. Energy Consumption by Office and
Telecommunications Equipment in Commercial Buildings Volume 1: Energy Consumption
Baseline Final Report. Office of Building Technology State and Community Programs. Contract
No.: DE-AC01-96CE23798.
Sanchez, Marla, Carrie Webber, Richard Brown, John Busch, Margaret Pinckard, Judy Roberson. 2007.
Space Heaters, Computers, Cell Phone Chargers: How Plugged In Are Commercial Buildings?
LBNL-62397.
U.S. Census Bureau. 2002. Economic Fact Sheet: California. Selected Statistics from the 2002 Economic
Census. Retrieved from
Office Plug Load Field Monitoring Report | December 2008 37
http://factfinder.census.gov/servlet/SAFFEconFacts?_event=&geo_id=04000US06&_geoContext
=01000US%7C04000US06&_street=&_county=&_cityTown=&_state=04000US06&_zip=&_lang=
en&_sse=on&ActiveGeoDiv=&_useEV=&pctxt=bg&pgsl=040&_submenuId=business_1&ds_nam
e=&_ci_nbr
Office Plug Load Field Monitoring Report | December 2008 38
Appendices
Methodology Design Matrix
Table 8. Study Design Prioritization
Considerations Priority Level Rationale
low high
Number of Participants Sample size needed for statistical significance too
large for project budget. Sample size designed to
include variety of offices.
Climate Zone Distribution Representation from both Northern and Southern
CA desired. Significant variations in plug load
energy use due to climate not expected.
Building Type Distribution Plug load energy use not expected to be heavily
dependent on building type.
Demographics of Employees Plug load energy use not expected to be heavily
dependent on employee demographics.
Business Types Targeted three business types in to investigate
plug load energy use variances among
businesses with different functions.
Number of Devices Metered Goal was to create as large a data set of office
plug loads as possible within the parameters of
the study.
Number of Days Metered Intent was to gather weekday and weekend data;
two weeks and two weekends would smooth out
any daily irregularities.
Interval of Meter Files Since a main goal of the study was to understand
how office plug loads are operated in the field, 1-
minute metering intervals would provide detailed
resolution on product operation.
Office Plug Load Field Monitoring Report | December 2008 39
Study Participant Data
Table 9. Study Participant Data
Business Type NAICS Square Footage
Full-Time Employees Surveyed Metered Sonoma San Diego
Legal Services 5411 5000 20 X X X
Legal Services 5411 38000 275 X X X
Legal Services 5411 X X
Legal Services 5411 3000 10 X X X
Legal Services 5411 800 3 X X
Legal Services 5411 6967 20 X X X
Legal Services 5411 3000 7 X X X
Legal Services 5411 13000 32 X X
Legal Services 5411 1500 X X
Legal Services 5411 3000 4 X X X
Accounting/Tax
preparation/etc. 5412 1300 7 X X
Accounting/Tax
preparation/etc. 5412 1100 3 X X
Accounting/Tax
preparation/etc. 5412 4389 24 X X
Accounting/Tax
preparation/etc. 5412 8000 26 X X X
Accounting/Tax
preparation/etc. 5412 1700 3 X X
Accounting/Tax
preparation/etc. 5412 750 2 X X X
Accounting/Tax
preparation/etc. 5412 1574 6 X X X
Architectural/Engineering 5413 1780 7 X X
Architectural/Engineering 5413 4200 16 X X
Architectural/Engineering 5413 6000 20 X X X
Architectural/Engineering 5413 800 2 X X X
Architectural/Engineering 5413 1000 4 X X X
Architectural/Engineering 5413 6500 16 X X X
Architectural/Engineering 5413 4200 14 X X X
Architectural/Engineering 5413 2000 16 X X
Architectural/Engineering 5413 4700 25 X X
Architectural/Engineering 5413 10000 45 X X
Architectural/Engineering 5413 4850 11 X X X
Architectural/Engineering 5413 1200 6 X X
Architectural/Engineering 5413 25000 100 X X X
Architectural/Engineering 5413 5500 16 X X
Computer Systems Design 5415 5000 12 X X X
Office Plug Load Field Monitoring Report | December 2008 40
Business Type NAICS Square Footage
Full-Time Employees Surveyed Metered Sonoma San Diego
Computer Systems Design 5415 3700 12 X X
Computer Systems Design 5415 1100 3 X X
Computer Systems Design 5415 7500 20 X X
Computer Systems Design 5415 1100 7 X X X
Computer Systems Design 5415 5000 16 X X X
Computer Systems Design 5415 3000 13 X X X
Computer Systems Design 5415 17000 48 X X
Computer Systems Design 5415 350 22 X X X
Computer Systems Design 5415 400 2 X X
Computer Systems Design 5415 6000 25 X X
Computer Systems Design 5415 1000 3 X X X
Computer Systems Design 5415 4983 6 X X X
Computer Systems Design 5415 1000 2 X X X
Computer Systems Design 5415 4650 13 X X
Computer Systems Design 5415 3324 9 X X X
Office Plug Load Field Monitoring Report | December 2008 41
Device List with Metering Prioritization
Table 10. Device Prioritization with Number of Devices Counted and Metered
Product Name Number of
Surveyed Items Number of
Metered Items Prioritization
Level
Computer, desktop 780 61 1
Computer, integrated-LCD 6 0 1
Computer, notebook 120 20 1
Computer display, CRT 203 21 1
Computer display, LCD 752 84 1
Projector, video 18 4 1
Copier 27 0 1
Multifunction device, inkjet 19 3 1
Multifunction device, laser 64 18 1
Printer, inkjet 71 13 1
Printer, laser 232 33 1
Printer, thermal 1 0 1
Printer, solid ink 0 0 1
Printer, wide format 21 7 1
Minicomputer/Thin client 44 6 1
External drive (CD, DVD) 58 2 1
Whiteboard, digital 2 0 1
Phone, switchboard 24 0 1
Audio minisystem 0 0 2
Charger, digital music player 18 0 2
Speakers, powered 73 6 2
Stereo, portable 46 1 2
Subwoofer 15 1 2
Charger, bar code scanner 0 0 2
Bar Code Scanner (no battery charger) 0 0 2
Television, LCD 7 2 2
Television, plasma 6 0 2
Television, rear projection 0 0 2
Television, CRT 21 0 2
Television/VCR Combination 2 0 2
Scale, digital 16 0 2
Fax, inkjet 8 0 2
Fax, laser 16 2 2
Fax, thermal 0 0 2
Mailing machine 28 0 2
Scanner, document 34 1 2
Scanner, flatbed 19 0 2
Scanner, slide 0 0 2
Office Plug Load Field Monitoring Report | December 2008 42
Product Name Number of
Surveyed Items Number of
Metered Items Prioritization
Level
Scanner, wide format 1 0 2
Amplifier, ethernet broadband
distribution 0 0 2
Hub or Switch, ethernet 134 9 2
Hub or Switch, USB 40 2 2
Firewall device 15 0 2
Modem, cable 3 0 2
Modem, DSL 8 0 2
Wireless access point 38 0 2
Charger, PDA 30 2 2
Speakers, computer 188 18 2
Tablet, pen (powered) 2 0 2
Charger, mobile phone 71 3 2
Dictation machine 0 0 2
Intercom 4 0 2
Phone, conference 21 0 2
Phone, corded (powered) 363 12 2
Phone, cordless 10 0 2
Phone, cordless with answering
machine 0 0 2
DVD player 3 1 2
DVD recorder 0 0 2
VCR 6 0 2
VCR/DVD 4 0 2
Adding, machine 143 12 2
Shredder 60 4 2
Time stamper 0 0 2
Mug warmer (powered) 9 0 2
Oven, microwave 45 2 2
Air cleaner, portable 0 0 2
Air conditioning, window mounted 0 0 2
Evaporative cooler, window mount 0 0 2
Fan, portable 152 6 2
Fan, window 0 0 2
Humidifier 0 0 2
Space heater, portable 44 4 2
Lights, holiday 0 0 2
Lamp, table 156 16 2
Lamp, floor 0 0 2
Lamp, desk attachment 236 9 2
Water dispenser, bottled 28 0 2
Charger, battery 67 2 2
Charger, cordless power tool 0 0 2
Office Plug Load Field Monitoring Report | December 2008 43
Product Name Number of
Surveyed Items Number of
Metered Items Prioritization
Level
Refrigerator, mini 39 3 2
Charger, smart phone 0 0 2
Amplifier 0 0 3
Cassette deck 0 0 3
CD player 13 3 3
CD player, portable 27 7 3
Equalizer (audio) 0 0 3
Radio, table 24 2 3
Receiver (audio) 6 0 3
Speakers, wireless (base station) 0 0 3
Speakers, wireless (speakers) 0 0 3
Computer, integrated-CRT 4 0 3
Game console, portable 0 0 3
Printer, impact (dot matrix and other) 0 0 3
Scanner, business card 1 0 3
Scanner, receipt (with external power
supply) 0 0 3
Modem, POTS 0 0 3
Modulator, audio/visual (powered) 0 0 3
Tape drive 4 0 3
Set-top box, digital cable 1 0 3
Set-top box, digital cable with PVR 0 0 3
Set-top box, game console 0 0 3
Set-top box, game console with internet 0 0 3
Set-top box, internet 0 0 3
Set-top box, PVR 0 0 3
Set-top box, satellite 0 0 3
Set-top box, satellite with PVR 0 0 3
Answering machine 1 1 3
Caller ID unit 0 0 3
Charger, still camera 2 0 3
Charger, video camera 0 0 3
Videocassette rewinder 0 0 3
Game console, commercial 0 0 3
Typewriter, Electric 24 1 3
Clock 4 1 3
Clock, radio 21 4 3
Coffee maker 50 10 3
Espresso maker, residential 4 1 3
Vacuum, rechargeable 9 0 3
Light box 3 0 3
Light, illuminated table 2 0 3
Charger, hedge trimmer 0 0 3
Office Plug Load Field Monitoring Report | December 2008 44
Product Name Number of
Surveyed Items Number of
Metered Items Prioritization
Level
Charger, weed trimmer 0 0 3
External power supply 4 0 3
Charger, wheelchair or golf cart 0 0 3
Charger, bicycle light 0 0 3
Tuner 0 0 3
Dock, notebook 77 4
Projector, slide 0 0
Printer, photo 0 0
Printer, receipt size (mini) 0 0
Router, Ethernet 11 0
Server, desktop-derived 79 0
Server, rack 54 0
Mainframe 0 0
Network equipment, IP telephone
adaptor 0 0
CD recorder 0 0
Security system 9 0
Set-top box, analog cable 0 0
Binding machine (electronic) 1 0
Hole punch (powered) 3 0
Laminator 1 0
Pencil sharpener 130 3
Stapler 52 2
Automatic griddles 2 0
Blender 6 0
Coffee grinder 5 2
Corn popper, air 0 0
Corn popper, hot oil 0 0
Hot plate (kitchen) 1 0
Kettle, electric 6 0
Toaster 16 0
Toaster oven 18 1
Vacuum, standard 19 0
Aquarium 4 0
Dehumidifier 0 0
Fan, range hood 0 0
Night light, interior 3 0
Timer, exterior (plug powered) 0 0
Timer, interior (plug powered) 1 0
Garbage disposal 2 0
Refrigerator, wine cooler 0 0
Trash compactor 0 0
Vending machine, cold 3 0
Office Plug Load Field Monitoring Report | December 2008 45
Product Name Number of
Surveyed Items Number of
Metered Items Prioritization
Level
Vending machine, hot 1 0
Vending machine, room temperature 1 0
Fountain, indoor 8 0
Air freshener (plug in) 4 0
Curling iron 0 0
Hair dryer 0 0
Home medical equipment 0 0
Water softener 0 0
Power strip 1027 0
Surge protector 0 0
Uninterruptible power supply, desktop 46 0
Uninterruptible power supply, server 70 0
Floor polisher 0 0
Power tool, corded 6 0
Clothes dryer, electric 0 0
Clothes dryer, gas 0 0
Clothes washer, horizontal axis 0 0
Clothes washer, standard 0 0
Cooktop, electric 0 0
Cooktop, gas 0 0
Dishwasher 4 0
Freezer 1 0
Oven, electric 0 0
Oven, gas 0 0
Refrigerator/Freezer 26 0
Other 149 0
Charger, miscellaneous 52 1
Office Plug Load Field Monitoring Report | December 2008 46
Determination of Power States
The power states for a product were determined using the following process. First, the data was put
through a 15-minute sliding average low pass filter. This has two effects: It averages out much of the
noise (which is usually about 0.2 W but can be 0.6 W or more for switch-mode power supplies at low
power levels). The sliding window also has the property that it remains constant only if the power level
stays constant for mode than 15 minutes. This was used to impose the requirement of a continuous
period. The second step was to identify those samples for which the sample value was the same as the
15-minute average (within 0.15 W). If the power level fluctuated, there would be occasional chance hits,
but steady power consumption would generate a large number of hits. The next step was to form a
histogram of the hits, using the data resolution of 0.1 W, and then scan the histogram for peaks. The
requirement is that it must have at least 40 hits, have more hits than the surrounding eight points, and
have more hits than the sum of the second, third, and fourth points on either side. This last requirement
means the peak must be narrower than about 0.2 W (half width at half maximum). These conditions
combined effectively impose the requirement of a narrow power range. The 40 hit minimum imposes the
requirement that the product spends a significant amount of time (at least 40 minutes) in each power
state.
This process was quite reliable for picking out power states. Power states that were used for only one
hour in the two week metering period were identified, yet the process was quite immune to both noise and
all the erratic behavior of products in active mode. If the active mode included stable power states, this
process did pick them out. Common examples were a battery charger whose power decreased as the
battery charged and eventually settled into a steady maintenance mode. This process identified the
maintenance mode separate from the charging mode. Another example was a computer that spent much
of its active time in the idle loop. When the computer was actually processing, the power consumption
rose briefly to a higher (and variable) level, and then dropped back to idle. This procedure picked out the
idle loop power state as distinct from the higher power level of computations.
Next, each sample was assigned to a power state according to the following criteria. A sample was
assigned to the ―disconnected‖ mode if it was one of at least 10 contiguous samples with zero power. A
reading (minute) was assigned to a power state if it was within 0.2 W of that power state. Power readings
were assigned to ―active‖ mode if they were above the highest stable power state. This left gaps of data
samples that were not yet assigned. Next, a data point which was immediately before or after a power
state and is within 0.4 W of that state is assigned to it. Finally, a gap that has the same state at both ends
and every point between is within 0.6 W of that state is assigned to the state. This left the very occasional
data point with substantial noise (which are too few to matter in the energy analysis) and those data that
are clearly between power states. These points were generally samples that were between two power
states in active mode. They were assigned to the next-higher state to ensure that they did not confuse a
standby or low-power mode.
Office Plug Load Field Monitoring Report | December 2008 47
Average Power by Mode
Table 11. Average Power Use by Mode
Product Name Number Active (W) Idle (W) Sleep (W) Standby (W)
Co
mp
ute
rs &
Mo
nit
ors
Co
mp
ute
rs Desktop Computer N=61 78.92 45.58 3.22 2.21
Notebook Computer N=20 74.72 30.33 1.56 1.59
Minicomputer/Thin client N=6 26.67 31.01 14.04 1.83
Notebook Dock N=4 26.29 1.11
Mo
nit
ors
CRT Display N=21 70.56 64.18 45.86 2.60
LCD Display N=84 34.24 26.43 6.19 0.88
Off
ice E
lectr
on
ics
Imag
ing
Laser MFD N=18 75.73 26.13 5.44 5.45
Inkjet MFD N=3 26.04 11.14 4.66
Laser Printer N=33 130.14 18.99 11.37
Inkjet Printer N=13 64.00 6.75 4.68 2.69
Wide Format Printer N=7 86.80 28.58 5.62
Document Scanner N=1 10.13 4.03
Laser Fax N=2 32.28 5.71
Co
mp
ute
r
Peri
ph
era
ls
Computer Speakers N=18 5.95 2.43 1.66
High Power Computer
Speakers N=2 73.05 1.35
External Drive N=2 28.43 10.73 .95
Ethernet Hub or Switch N=9 16.96 7.97 5.87 1.29
USB Hub or Switch N=2 26.03 14.05 5.92 0.56
Office Plug Load Field Monitoring Report | December 2008 48
Product Name Number Active (W) Idle (W) Sleep (W) Standby (W)
Mis
cellan
eo
us
Au
dio
/ V
isu
al
Television, LCD N=2 58.17 3.14
DVD player N=1 1.28
Video Projector N=4 181.9 9.76 4.56
CD player N=3 8.26 2.06
Portable CD player N=7 17.95 2.95 1.27
Speakers N=6 32 10 3 1
Portable Stereo N=1 7.5 3.31 0.88
Subwoofer N=1 0 6.96
Table Radio N=2 2.78 1.4
Tele
ph
on
y Charger, mobile phone N=3 2.12 0.65 0.15
Phone, corded (powered) N=12 4.85 2.43
Bu
sin
es
s E
qu
ipm
en
t
Adding, machine N=12 3.58 3.57 1.58
Battery charger N=2 3.37 0 1.26
Clock N=1 1.4
Clock, radio N=4 5.37 2.99 4.04
Coffee grinder N=2 1.25 0.21
Coffee Maker N=10 464.01 40.25 1.77
Espresso maker, residential N=1 369.38 2.24
Fan, portable N=2 0.63
Lamp, desk attachment N=9 35.35 23.21 0.57
Lamp, table N=16 41.7 13.38 0.91
Shredder N=4 78.36 0.77
Space heater, portable N=4 937.65 1.03
Stapler N=2 1.73 0.81 1.22
Toaster oven N=1 1057.9 0.03
Typewriter, Electric N=1 7.13 3.38
Office Plug Load Field Monitoring Report | December 2008 49
Average Annual Energy Year per Mode per Device
Table 12. Average Annual Energy Year per Mode per Device
Product Name Number Active (kWh) Idle (kWh) Sleep (kWh) Standby (kWh)
Co
mp
ute
rs &
Mo
nit
ors
Co
mp
ute
rs Desktop Computer N=61 221.11 171.86 2.30 11.81
Notebook Computer N=20 41.88 45.11 1.83 7.41
Minicomputer/Thin client N=6 56.05 191.29 18.19 7.15
Notebook Dock N=4 8.99 0 0 8.46
Mo
nit
ors
CRT Display N=21 130.95 42.53 32.62 13.85
LCD Display N=84 58.42 59.42 8.80 5.54
Off
ice E
lectr
on
ics
Imag
ing
Laser MFD N=18 111.84 49.28 0.69 35.16
Inkjet MFD N=3 1.60 6.74 0 39.10
Laser Printer N=33 97.97 93.54 0 88.29
Inkjet Printer N=13 26.27 7.72 28.79 21.15
Wide Format Printer N=7 128.14 146.70 0 46.90
Document Scanner N=1 2.41 0 0 34.64
Laser Fax N=2 14.82 0 0 48.36
Co
mp
ute
r
Peri
ph
era
ls
Computer Speakers N=18 13.82 20.91 0 11.42
High Power Computer
Speakers N=2 119.89 0 0 1.08
External Drive N=2 49.79 0 7.22 8.16
Ethernet Hub or Switch N=9 48.52 33.42 1.00 10.75
USB Hub or Switch N=2 9.83 27.39 4.06 3.86
Office Plug Load Field Monitoring Report | December 2008 50
Product Name Number Active (kWh) Idle (kWh) Sleep (kWh) Standby (kWh)
Mis
cellan
eo
us
Au
dio
/ V
isu
al
Television, LCD N=2 161.67 0 0 19.01
DVD player N=1 0 0 0 11.31
Video Projector N=4 25.24 0 17.12 34.65
CD player N=3 2.50 0 0 17.26
Portable CD player N=7 18.79 5.41 0 8.34
Speakers N=6 18.91 29.36 2.45 5.88
Portable Stereo N=1 2.16 3.23 0 6.66
Subwoofer N=1 0 61.58 0 0
Table Radio N=2 7.83 0 0 9.60
Tele
ph
on
y
Charger, mobile phone N=3 0.45 0.04 0 1.08
Phone, corded (powered) N=12 0.34 19.63 0 0
Bu
sin
es
s E
qu
ipm
en
t
Adding, machine N=12 7.34 24.73 0 13.54
Battery charger N=2 0.57 0 0 10.91
Clock N=1 0 12.41 0 0.00
Clock, radio N=4 4.26 25.38 0 17.95
Coffee grinder N=2 0.00 10.01 0 0.93
Coffee maker N=10 490.66 276.25 0 11.79
Espresso maker,
residential N=1 153.75 0 0 18.91
Fan, portable N=2 0 0 0 5.56
Lamp, desk attachment N=9 33.73 14.19 0 4.19
Lamp, table N=16 78.19 27.48 0 2.65
Shredder N=4 165.69 0 0 2.38
Space heater, portable N=4 108.84 0 0 0.16
Stapler N=2 0.01 6.83 0 10.75
Toaster oven N=1 49.65 0 0 0.08
Typewriter, Electric N=1 0.04 0 0 29.68
Office Plug Load Field Monitoring Report | December 2008 51
Miscellaneous Category Duty Cycle Data
Table 13. Miscellaneous Category Duty Cycle Data
Product
Number
Metered
Average
Time in
Active
Average
Time in
Idle
Average
Time in
Sleep
Average
Time in
Standby
Average
Time in
Disconnect
Television, LCD 2 16% 0% 0% 34% 50%
DVD player 1 0% 0% 0% 100% 0%
Video Projector 4 6% 0% 5% 64% 25%
CD player 3 5% 0% 0% 95% 0%
Portable CD Player 7 5% 9% 0% 75% 11%
Speakers 6 7% 37% 2% 40% 14%
Portable Stereo 1 3% 11% 0% 86% 0%
Subwoofer 1 0% 100% 0% 0% 0%
Table Radio 2 16% 0% 0% 84% 0%
Mobile Phone Charger 3 2% .25% 0% 33.75% 64%
Corded Phone 11 0% 92% 0% 0% 8%
Answering Machine 2 0% 100% 0% 0% 0%
Adding Machine 12 5% 54% 0% 25% 16%
Battery Charger 2 1% 0% 0% 49% 50%
Clock 1 0% 100% 0% 0% 0%
Clock Radio 4 6% 72% 0% 22% 0%
Coffee Maker 10 25.5% 16% 0% 46% 12.5%
Coffee Grinder 2 0% 45% 0% 25% 30%
Espresso Maker 1 5% 0% 0% 95% 0%
Lamp, Desk Attachment 9 20% 1% 0% 9% 70%
Lamp, Table 16 14.5% 1.5% 0% 4% 80%
Shredder 4 25% 0% 0% 43% 32%
Space heater, portable 4 1% 0% 0% 99% 0%
Stapler 2 0% 48% 0% 50% 2%
Toaster Oven 1 1% 0% 0% 33% 66%
Typewriter, Electric 1 0% 0% 0% 100% 0%
PDA Charger 2 1% 44% 0% 50% 5%